Recurrent Neural Network to Analyze the Reiner–Philippoff Fluidic System in Carbon Nanotubes Along Stretchable Riga Plate

ABSTRACT The present research investigates the impact of carbon nanotubes on fluid dynamics in order to enhance heat transmission and stabilize the moving base fluid in contemporary technology. The Reiner–Philippoff fluid model has pseudo‐plastic, dilatant and non‐Newtonian behavior with variable viscosity, facilitating the transition of the fluid between different rheological states. This study aims to employ a recurrent neural network with a Bayesian regularization optimizer (RNN‐BRO) to investigate the feasibility of using single and multi‐wall carbon nanotubes in the flow of Reiner–Philippoff fluid (RPF‐CNTs) under magnetohydrodynamic conditions along a stretching sheet. The synthetic data for the RPF‐CNTs model is generated by employing Adams numerical method across various parameter settings of the fluid flow, fluid temperature, and nanoparticle concentration gradients. The resultant data set was used as the testing, training, and validation sets for the proposed RNN‐BRO. Furthermore, the effects of various physical parameters on the flow behavior, fluid temperature and concentration profile of RPF‐CNTs are analyzed. The results revealed that an increase in magnetic parameter results in a greater boundary layer thickness of momentum in both SWCNT and MWCNT; however, the reduction in flow resistance is more pronounced in SWCNT compared to MWCNT with a boost in the effective Prandtl number. The performance and efficiency of RNN‐BRO were assessed using criteria such as mean square error, regression studies, analysis of the error histograms, mu, gradients, and by the absolute error ranging from 10 −04 to 10 −12 , which indicated the efficacy of the proposed RNN‐BRO technique.